地震研究

针对如何选取合适的影响因素进行砖木结构房屋地震破坏合理评估的问题,提出了一种基于主成分分析与BP神经网络相融合的云南砖木结构房屋地震破坏评估方法,通过灰色关联度模型剔除对砖木结构房屋发生地震破坏影响较小的因素得到关键因子,采用主成分分析法从关键因子中提取主要成分,最后利用BP神经网络模型对处理后的主要成分进行训练,建立砖木结构房屋地震破坏比例预测模型,并利用实际震例进行验证。结果表明:本文方法相较于传统脆弱性曲线拟合方法和BP神经网络模型,其预测的砖木结构房屋地震破坏比例的预测精度更高、普适性更好。

There are a lot of factors affecting the earthquake damage to brick-timber houses,and selecting appropriate influencing factors is an important guarantee for an accurate and reasonable assessment of the earthquake damage to masonry-timber houses. The question is that when using the traditional methods,it is difficult to choose proper factors. In this paper,a method for assessing the earthquake damage to masonry-timber houses in Yunnan is proposed by integrating the principal component analysis and the neural network. Firstly,the less-influencing factors are eliminated through the gray correlation degree model and the key factors are obtained. Secondly,the main components from the key factors are extracted through the principal component analysis.Finally,the main components are trained through the BP neural network model,and a pre-estimating model for the earthquake-damage ratio of masonry-limber houses is established. This method is tested by using the data from the post-earthquake investigation of some historical earthquakes,and the results show that this method is more accurate and applicable for pre-estimating the earthquake damage ratio of masonry-timber houses than the traditional vulnerability curve fitting and the neural network model.

引言
1 数据与方法
2 研究结果
3 结论

表1 砖木结构房屋损伤评估训练样本<br/>Tab.1 Training samples for assessing the damaged masonry-timber houses

表1 砖木结构房屋损伤评估训练样本
Tab.1 Training samples for assessing the damaged masonry-timber houses

表2 砖木结构房屋损伤评估测试样本<br/>Tab.2 Testing samples for assessing the damaged masonry-timber houses

表2 砖木结构房屋损伤评估测试样本
Tab.2 Testing samples for assessing the damaged masonry-timber houses

图1 基于PCA与BP神经网络相融合的砖木结构房屋地震破坏评估方法技术路线<br/>Fig.1 Technical route of the assessment of earthquake-damaged masonry-limber houses by integrating PCA with BP neural network

图1 基于PCA与BP神经网络相融合的砖木结构房屋地震破坏评估方法技术路线
Fig.1 Technical route of the assessment of earthquake-damaged masonry-limber houses by integrating PCA with BP neural network

图2 各影响因子灰色关联度矩阵<br/>Fig.2 Gray correlation matrix of each impact factor

图2 各影响因子灰色关联度矩阵
Fig.2 Gray correlation matrix of each impact factor

图3 影响因子关联度降序排列<br/>Fig.3 Impact-factor correlations in descending order

图3 影响因子关联度降序排列
Fig.3 Impact-factor correlations in descending order

表3 主成分特征值及贡献率<br/>Tab.3 Eigenvalues and contribution rates of the principal components

表3 主成分特征值及贡献率
Tab.3 Eigenvalues and contribution rates of the principal components

图4 关键因子得分系数矩阵<br/>Fig.4 Score coefficient matrix of the key factors

图4 关键因子得分系数矩阵
Fig.4 Score coefficient matrix of the key factors

图5 BP神经网络模型结构图<br/>Fig.5 Structure of the BP neural network model

图5 BP神经网络模型结构图
Fig.5 Structure of the BP neural network model

图6 砖木结构房屋地震破坏脆弱性拟合曲线<br/>Fig.6 Vulnerability fitting curve of the masonry-timber houses

图6 砖木结构房屋地震破坏脆弱性拟合曲线
Fig.6 Vulnerability fitting curve of the masonry-timber houses

表4 预测结果对比分析<br/>Tab.4 Comparison and analysis of the pre-estimated results

表4 预测结果对比分析
Tab.4 Comparison and analysis of the pre-estimated results

图7 3种方法对砖木结构房屋地震破坏比预测结果对比分析<br/>Fig.7 Comparison pre-estimated results of the damage ratio of the masonry-timber houses by three methods

图7 3种方法对砖木结构房屋地震破坏比预测结果对比分析
Fig.7 Comparison pre-estimated results of the damage ratio of the masonry-timber houses by three methods

成小平,胡聿贤,帅向华.2000.基于神经网络模型的房屋震害易损性估计方法[J].自然灾害学报,9(2):68-73.
杜浩国,余庆坤,陈亮,等.2019.云南红河县城区房屋地震灾害三维仿真模拟方法[J].地震研究,42(2):218-229.
侯林锋,徐超,石树中,等.2020.城市建筑群地震灾害风险评估技术探索及应用[J].华北地震科学,38(3):69-73.
胡进军,张辉,靳超越,等.2021.基于PCA及PSO智能算法的地震动合成方法——以中国西部中强地震为例[J].工程力学,38(3):159-168.
黄佩蒂.2018.境外地震人员伤亡影响因素研究[D].哈尔滨:中国地震局工程力学研究所.
黄文涛,熊永良,张莹,等.2022.川南地区民居建筑地震易损性研究[J].华北地震科学,40(4):40-49.
贾婧,窦圣宇,范国玺,等.2020.基于熵权法和灰色关联分析法的海岛地震应急能力评价研究[J].世界地震工程,36(3):233-241.
蒋复量,吴浩楠,刘永,等.2020.基于灰色关联度和变权理论的地震条件下铀尾矿库群坝安全评价研究[J].南华大学学报:自然科学版,34(1):22-28.
金赟赟,李杰.2020.基于易损性贝叶斯网络的群体建筑快速震害预测[J].自然灾害学报,29(5):64-72.
李桥,巨能攀,黄健,等.2020.基于主成分分析的PSO-ELM Adaboost算法耦合模型在极震区泥石流物源动储量计算中的应用[J].科学技术与工程,20(15):5961-5970.
刘晶晶,宁宝坤,吕瑞瑞,等.2017.震后典型建筑物倒塌分类及救援特点分析[J].震灾防御技术,12(1):220-229.
刘毅,吴绍洪,徐中春,等.2011.自然灾害风险评估与分级方法论探研——以山西省地震灾害风险为例[J].地理研究,30(2):195-208.
刘育林,周爱红,袁颖.2022.基于GRA-PCA-BP神经网络模型的单沟泥石流危险性预测[J].河北地质大学学报,45(4):44-51.
施唯,王东明.2022.基于BP神经网络的有限特征参数砌体结构震害等级推演[J].世界地震工程,38(2):160-168.
施伟华,陈坤华,杨树明,等.2011.盈江2011年MS5.8与2008年MS5.9地震的震害差异及原因[J].地震研究,34(4):518-524.
孙柏涛,张桂欣.2012.汶川8.0级地震中各类建筑结构地震易损性统计分析[J].土木工程学报,45(5):26-30.
汤皓,陈国兴.2006.基于灰关联与人工神经网络综合评价模型的多层砖房震害预测[J].世界地震工程,22(4):133-139.
王晨晖,袁颖,刘立申,等.2022.基于主成分分析法优化广义回归神经网络的地震震级预测[J].科学技术与工程,22(29):12733-12738.
杨钦杰,李蕾,梁结,等.2021.钦州市建筑物抗震性能普查及震害预测分析[J].华南地震,41(2):76-82.
臧石磊,冯石,李玉森,等.2019.基于模糊综合评判法的辽宁省老旧砌体房屋震害预测[J].地震研究,42(2):166-172.
张灿,陈文凯,林旭川.2022.基于地震动参数的城市建筑物震害模拟——以兰州市城关区为例[J].地震工程学报,44(1):172-182.
张方浩,卢永坤,邓树荣,等.2020.云南地区区域地震灾害特征[J].地震研究,43(1):134-143.
张雪华.2017.基于无人机影像的面向对象建筑物震害提取研究[D].哈尔滨:中国地震局地震预测研究所.
赵东升,靳京,吴绍洪.2022.区域地震灾害房屋脆弱性曲线构建研究[J].自然灾害学报,31(1):127-136.
钟江荣,张令心,沈俊凯.2021.基于震害资料的云南地区砌体结构易损性研究[J].华南地震,41(1):108-116.
周光全,非明伦,施伟华.2006.1992—2005年云南地震灾害损失与主要经济指标研究[J].地震研究,29(2):198-202.
周光全.2012.云南地震灾害损失评估及研究[M].昆明:云南科技出版社.
Tien J M.2017.Book review:Grey data analysis:Methods,models and applications[J].The Journal of Grey System,29(1):151-152.
Xu J,Chen Y,Xie T,et al.2019.Prediction of triaxial behavior of recycled aggregate concrete using multivariable regression and artificial neural network techniques[J].Construction and Building Materials,226(30):534-554.

备注

引言

1 数据与方法

2 研究结果

3 结论

期刊信息